Trying to create a visualization for embeddings

API
1

I’m trying to build a text classification model using the Openai API - but I don’t know how to make a JSONL file that could be read by it. unfortunately no JSONL file was provided so I can’t base my work off of one, If anyone could show me how to create a file that could be read by the embeddings endpoint and the visualization code, I would be incredibly appreciative. Thank you for your time!

2 Likes
2

I’m so excited to see what comes from this.

1 Like
4

Hi @curranjanssens

Here’s a great explainer on the JSONL files on the OpenAI docs.

5

image
image1030×222 23.6 KB

Given this piece of code from the link: (https://github.com/openai/openai-python/blob/main/examples/embeddings/Visualize_in_3d.ipynb ), what part of the JSONL file would be where “category” would be

7

Thanks! that was my guess but I was a little lost, really appreciate your help

2 Likes
8

Hi,
I stumbled upon the same problem and I was wondering if you managed to create something? I am trying to visualise the embeddings in 2D and 3D space in the browser by uploading a JSON file.

And also, do you think T-SNE is better for visualisation of the ada embeddings (without reducing the dimensions beforehand with something like PCA)?